This proposal addresses the major challenge of improving health outcomes for children with cancer and other complex conditions, for whom the effectiveness of outpatient care depends on coordination across a diverse group of caregivers that includes parents, community support and pediatric care providers. While a goal-centered care plan is considered a critical tool for coordinating care, such care plans have not been effectively integrated into the pediatric care system, and current health information technology provides insufficient support for their use. The proposed fundamental research on information-sharing algorithms aims to provide foundations for intelligent interactive systems that assist care teams in using integrated care plans in the context of shared decision-making to improve the care of children with complex chronic conditions. The primary scientific computational aim is to develop new multi-agent systems (MAS) algorithms that are able to reason about plans with only partial information, in conjunction with human-computer interaction methods to support use of integrated care plans. To meet this aim, a new representation, mutual influence potential networks, will be developed. The primary health science aims are to identify appropriate vocabularies for patient-provider communication about care goals and to assess the potential for intelligent information sharing capabilities to improve meaningful use of care plans. The primary technological aim is to develop GoalKeeper, a system for supporting care plan use. The project is a collaborative effort of computer scientists at Harvard University, who bring expertise in teamwork models, collaborative human-computer interfaces and personally adaptive human-computer interaction, and physician scientists at Stanford University, who bring expertise in health-communication, health behavior, and health-services. This project will make fundamental contributions to computer science through the development of novel MAS representations and algorithms. It will contribute to health sciences by providing clear definitions and effective strategies for shared goal setting and by assessing the potential of technological support for care coordination to improve health outcomes for children with cancer and other complex chronic conditions.